Posts

Large language models enable fracking of documents. Historically, extracting value from unstructured text files has been difficult. But LLMs do this beautifully, pumping value from one of the hardest places to mine.

We have a collection of thousands of notes researching startups. We are tinkering with deploying large language models on top of them.

Here are some quick observations about our initial experiments :

The Future is Constellations of Models. When faced with a search box, a user might ask quantitative questions. For example, how many people from Google have a met in the last month?

In There’s No Such Thing as Series A Metrics, Charles Hudson explains that there is no magic milestone to raise a Series A.

In this environment, I agree. The $1m ARR figure used to hold in 2018 & early 2019. But the data shows how much the market differs from a few years ago.

Series A round size standard deviation has grown by between 4-5x in 4 years.

Threads, Facebook’s newest social product & direct Twitter competitor, has surged to 70m monthly active users (MAU) as of this morning.

What does that mean for B2B marketers?

About 1350m users open the Instagram app every month. About 368m click on Twitter’s blue bird, 72% fewer.

Threads, launched on July 5th, counts about 70m or about 20% the size of Twitter. At 70m users, Threads current population is 5% of Instagram’s.

In August of last year, public markets switched from valuing SaaS companies primarily on growth to primarily on profitability. In the past two months, high-flying software companies have seen 40-112% increases in valuation.

I wondered if the public markets have begun to change the way they value these businesses.

The answer is yes - if you’re a top quartile company.

Yesterday, Databricks announced its intent to acquire Mosaic for $1.3b. Perhaps not coincidentally, Snowflake announced a deepened partnership with Nvidia to offer customers models & training on Nvidia’s Nemo platform.

Clouds are picking teams in one of the most important dislocations in software.

Cloud <- -> LLM Infrastructure
Microsoft <- -> OpenAI
Snowflake <- -> Nvidia
Databricks <- -> Mosaic
Google <- -> Anthropic
Oracle <- -> Cohere
Amazon <- -> HuggingFace

MotherDuck announced their product, managed DuckDB in the cloud, is available by invitation.

I have a raft of extra early invitations that I’ll distribute randomly on Monday. If you’d like to put your email in for a chance to receive one, just fill out this form.

DuckDB is one of the most powerful, lightweight, & fastest adopted analytical databases. MotherDuck extends DuckDB to the cloud. The video above with a french-accented Duck walks through the product with elan.

I think of marketing teams as hedge funds. Marketing teams develop a portfolio of different strategies to acquire leads.

Some days, content marketing works. A post challenging like this one hits HackerNews or a journalist covers the company.

Other days when content strategies struggle, a witty paid ad campaign entices some clicks & form fills. On some rainy Tuesday, the weekly webinar attracts an unusually promising audience.

Like an investment portfolio, any individual strategy may thrive one day & suffer the next. But as the effort broadens, leads compound at a steadier & steadier rate.

If a company adds AI to their pitch or nabs a .ai domain, will they raise money at a higher valuation?

Nvidia’s sudden surge to $1t in market capitalization in the public market suggests it could be the case.

The data tells a different story. I plotted the 75th percentile valuation of all US software companies by stage compared to AI companies over the last twelve years.

In the last month, Hubspot launched ChatSpot. Zendesk AI automates customer support. Salesforce launched AI Cloud. Adobe’s released Firefly. What do all of these launches & many other similar ones have in common?

The largest SaaS companies in the world have released English-language APIs. Anyone can write English to enter some input (e.g., “What are my top leads for this week?”) & receive an answer.

Previous APIs require learning programming languages to use & have been reserved for developers.

In the mid-2010s, every web1 company became a web2 company. This time it’s different. Many web3 companies will become web2 companies, too.

Why will web3 companies dress like SaaS mutton?

If the current pace continues, web3 startup fundraising will fall by 73% in 2023. The absence of those dollars will flatten & shrink the already modest addressable market for web3 software & infrastructure.